Welcome to DU!
The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards.
Join the community:
Create a free account
Support DU (and get rid of ads!):
Become a Star Member
Latest Breaking News
Editorials & Other Articles
General Discussion
The DU Lounge
All Forums
Issue Forums
Culture Forums
Alliance Forums
Region Forums
Support Forums
Help & Search
Reddit is people! Which means its search might not be so damaged by AI slop
Reddit has found that trafficking in human-authored content pays well in the AI era.
And the popular internet community website sees an opportunity to make inroads in the search business, even as web search is being swallowed by generative AI.
Company shares surged this week after Reddit reported $500 million in revenue for Q2 2025, a 78 percent increase. Net income reached $89 million, a welcome turnaround from a $10 million loss during Q2 2024.
CEO Steve Huffman said in a statement, "Reddit is built for this moment. In a world where connection is increasingly rare, our communities show how valuable human conversation and knowledge really are."
Reddit's fortunes in part can be attributed to demand from AI firms for the community site's historically human-authored posts, which they use for model training. The company's data licensing business, part of its "other revenue" line item, reached $35 million, up 24 percent year-over-year. Reddit last year struck content partnerships with OpenAI and Google. And in June this year, it sued Anthropic to halt uncompensated content scraping.
Huffman, during the company's conference call for investors, acknowledged the value of content licensing along with its limits. "What we've learned since we did those deals is that the Reddit Corpus is not essential, but appears to be growing in importance to some of the largest players on the Internet," he said.
Huffman expects Reddit will continue to learn how to structure its partnerships while also striving to build the best possible products on top of the Reddit Corpus – the company's content dataset.
And the popular internet community website sees an opportunity to make inroads in the search business, even as web search is being swallowed by generative AI.
Company shares surged this week after Reddit reported $500 million in revenue for Q2 2025, a 78 percent increase. Net income reached $89 million, a welcome turnaround from a $10 million loss during Q2 2024.
CEO Steve Huffman said in a statement, "Reddit is built for this moment. In a world where connection is increasingly rare, our communities show how valuable human conversation and knowledge really are."
Reddit's fortunes in part can be attributed to demand from AI firms for the community site's historically human-authored posts, which they use for model training. The company's data licensing business, part of its "other revenue" line item, reached $35 million, up 24 percent year-over-year. Reddit last year struck content partnerships with OpenAI and Google. And in June this year, it sued Anthropic to halt uncompensated content scraping.
Huffman, during the company's conference call for investors, acknowledged the value of content licensing along with its limits. "What we've learned since we did those deals is that the Reddit Corpus is not essential, but appears to be growing in importance to some of the largest players on the Internet," he said.
Huffman expects Reddit will continue to learn how to structure its partnerships while also striving to build the best possible products on top of the Reddit Corpus – the company's content dataset.
https://www.theregister.com/2025/08/02/reddit_ready_for_search_business/
1 replies
= new reply since forum marked as read
= new reply since forum marked as read
